Viewing finite-dimensional representations through infinite-dimensional ones (Q1306243)

Given a finite-dimensional algebra \(\Lambda\) and \({\mathcal U}\) a full subcategory of \(\Lambda\)-mod, \({\mathcal U}\) is said to be contravariantly finite in \(\Lambda\)-mod if each object in \(\Lambda\)-mod has an \({\mathcal U}\)-approximation. Recall that \(M\in \Lambda\)-mod has a (right) \({\mathcal U}\)-approximation if there is a homomorphism \(\phi: A\to M\) with the property that the induced map of functors \(\text{Hom}(-,\phi):\text{Hom}_{\Lambda}(-, A)|_{{\mathcal U}}\to \text{Hom}_{\Lambda}(-, M)|_{{\mathcal U}}\) is surjective. The present work is twofold. One of these is to describe criteria to indicate when the contravariant finiteness of \({\mathcal U}\) fails, through relative approximations. The second is to give an important and general result that the failure of contravariant finiteness of \({\mathcal U}\) reveals the existence of modules of infinite \(K\)-dimension. These modules in some sense take over the role which is played by the minimal right \({\mathcal U}\)-approximation in the contravariantly finite case. The authors call such special modules ``\({\mathcal U}\)-phantoms'', also using the idea of relative approximations. Furthermore, they present in the article many interesting and nice examples about the failure of contravariant finiteness of \({\mathcal P}^{\infty}(\Lambda\)-mod), exhibiting an infinite-dimensional \({\mathcal P}^{\infty}(\Lambda\)-mod)-phantom of some simple \(\Lambda\)-mod.